Vulnerability of Shallow Groundwater and Drinking-Water Wells to

Oct 27, 2006 - Seo Jin Ki , Chittaranjan Ray , Mohamed M. Hantush ... Kourakos , Thomas Harter. Environmental Modelling & Software 2014 52, 207-221 ...
0 downloads 0 Views 507KB Size
Environ. Sci. Technol. 2006, 40, 7834-7840

Vulnerability of Shallow Groundwater and Drinking-Water Wells to Nitrate in the United States BERNARD T. NOLAN* AND KERIE J. HITT U.S. Geological Survey, 413 National Center, Reston, Virginia 20192

Two nonlinear models were developed at the national scale to (1) predict contamination of shallow ground water (typically < 5 m deep) by nitrate from nonpoint sources and (2) to predict ambient nitrate concentration in deeper supplies used for drinking. The new models have several advantages over previous national-scale approaches. First, they predict nitrate concentration (rather than probability of occurrence), which can be directly compared with waterquality criteria. Second, the models share a mechanistic structure that segregates nitrogen (N) sources and physical factors that enhance or restrict nitrate transport and accumulation in ground water. Finally, data were spatially averaged to minimize small-scale variability so that the largescale influences of N loading, climate, and aquifer characteristics could more readily be identified. Results indicate that areas with high N application, high water input, well-drained soils, fractured rocks or those with high effective porosity, and lack of attenuation processes have the highest predicted nitrate concentration. The shallow groundwater model (mean square error or MSE ) 2.96) yielded a coefficient of determination (R2) of 0.801, indicating that much of the variation in nitrate concentration is explained by the model. Moderate to severe nitrate contamination is predicted to occur in the High Plains, northern Midwest, and selected other areas. The drinking-water model performed comparably (MSE ) 2.00, R2 ) 0.767) and predicts that the number of users on private wells and residing in moderately contaminated areas (>5 to e10 mg/L nitrate) decreases by 12% when simulation depth increases from 10 to 50 m.

Introduction Groundwater is an important national resource that provides drinking water for nearly half the people in the United States (U.S.). Unfortunately, the resource is susceptible to contamination by chemicals derived from the land surface. Nitrate is considered the most widespread contaminant in groundwater (1). Because nitrate is both soluble and mobile, it is prone to leaching through soils with infiltrating water. High nitrate concentration in groundwater is a human health concern. Prevention of methemoglobinemia in infants is the basis for the maximum contaminant level of 10 mg/L nitrate as N, established by the U.S. Environmental Protection Agency (1). Recent studies have associated nitrate in drinking * Corresponding author phone: +33 238 64 47 50; fax: +33 238 64 34 46; e-mail: [email protected]. 7834 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 24, 2006

water with several types of cancer (2-5). The causal role of nitrate is not conclusive because there are few such studies for a given type of cancer, and because it is difficult to evaluate the combined effect of nitrate intake from food and water. Nevertheless, the results are a cause for concern because the adverse effects are associated with nitrate concentrations as low as 2.5 mg/L (5). Relative background concentration of nitrate in shallow groundwaters of the U.S. is about 1 mg/L (6). Protection of drinking-water sources is a national priority and is mandated by the Safe Drinking Water Act (7). The U.S. Geological Survey’s (USGS’s) National Water-Quality Assessment (NAWQA) Program effectively monitors the occurrence and distribution of nitrate and other contaminants in groundwater and streams, using consistent sampling and analytical methods (8). It is impractical to monitor everywhere, however, so the availability of high-quality data is limited nationally. Data gaps can be addressed with regional and national water-quality models that use detailed spatial data on chemical loadings and environmental characteristics. Emperical models, in particular logistic regression, have been successfully applied at a variety of scales to predict the likelihood of contamination by various chemicals (9-24). Predicted probabilities, however, cannot be directly compared with water-quality standards or other concentrations relevant to human health, and predicting concentrations is problematic because water-quality data commonly are censored at the analytical reporting limit. Use of ordinary least-squares with simple substitution of values (e.g., zero or half the reporting limit) for censored data is inappropriate because the resulting model coefficients depend on the assumed values (25). The goal of the current study is to predict groundwater vulnerability to nitrate at the national scale, to complement measured data. Specific objectives are to reliably predict nitrate concentration in shallow groundwater and in that used for drinking and to describe the uncertainty of the predictions. We present a nonlinear approach to nationalscale Ground-WAter Vulnerability Assessment (GWAVA), which uses average characteristics of NAWQA monitoring networks. Use of network averages smoothes local variability that can obscure large-scale trends in nitrate concentration. By focusing on large-scale variability, the predominant processes influencing nitrate contamination at the national scale can more readily be identified. Compared with simple linear approaches, the model has a mechanistic structure with components representing nitrogen (N) sources, transport of nitrate to aquifers, and attenuation of nitrate in groundwater. Nonlinear regression models previously were developed at large spatial scales to predict N yields and phosphorus concentrations in streams of U.S. watersheds (26, 27). The SPAtially Referenced Regressions On Watersheds (SPARROW) model relates measured chemical transport rates in streams to spatial data comprising chemical source terms, land-towater delivery factors, and in-stream-decay factors. The SPARROW model reliably predicted contaminant transport in U.S. watersheds; the mean square error (MSE) of the N model was 0.45, and the coefficient of determination (R2) was 0.87. Two nonlinear models are presented in the current study, to predict nitrate concentration in U.S. groundwaters. The first (GWAVA-S) predicts nitrate contamination of shallow, recently recharged groundwater, which may or may not be used for drinking. The calibration data set represents 10.1021/es060911u Not subject to U.S. copyright. Publ. 2006 Am. Chem.Soc. Published on Web 10/27/2006

TABLE 1. Parameters of Nonlinear Regression Model for Nitrate in Shallow Groundwater (GWAVA-S) model parameter

units

estimated coefficient Nitrogen Source (β) 0.2265 0.4049 1.9600 0.006658 0.1473

standard error

significance level (p)

0.0722 0.1355 0.8466 0.0014 0.0585

0.0024 0.0037 0.0231